Re: libstdc++/2568

Re: libstdc++/2568

[Rainer Orth]

The following reply was made to PR libstdc++/2568; it has been noted by GNATS.

From: Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
To: Brad Lucier <lucier@math.purdue.edu>
Cc: bkoz@gcc.gnu.org, gcc-gnats@gcc.gnu.org, David.Billinghurst@riotinto.com,
        neil@gcc.gnu.org, amacleod@cygnus.com (Andrew Macleod)
Subject: Re: libstdc++/2568
Date: Fri, 25 May 2001 21:42:44 +0200 (MEST)

 Brad Lucier writes:
 
 > > * a sparc-sun-solaris2.8 configuration (sparcv7),
 > 
 > I just did one of these today, and I started a sparcv9-sun-solaris2.8
 > with (I believe) multilib support that should finish in a few
 
 sparc-sun-solaris2.8 is bootstrap and check are finished,
 sparcv9-sun-solaris2.8 bootstrap is done, check is well along, I'm waiting
 for this to finish before I start the -m32 check.
 
 > hours.  (The bootstrap is fairly fast, it's the make check
 > that kills me.)
 
 Bootstrap is generally ok on an 2 x 400 MHz E250, although the bootstrap
 takes considerably longer for sparcv9 (even up to the make compare), It's
 obvious that building all libraries twice takes longer :-)
 
 	Rainer
 
 -----------------------------------------------------------------------------
 Rainer Orth, Faculty of Technology, Bielefeld University
 
 Email: ro@TechFak.Uni-Bielefeld.DE

Re: libstdc++/2568

[Rainer Orth]

The following reply was made to PR libstdc++/2568; it has been noted by GNATS.

From: Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
To: Brad Lucier <lucier@math.purdue.edu>
Cc: bkoz@gcc.gnu.org, gcc-gnats@gcc.gnu.org, David.Billinghurst@riotinto.com,
        neil@gcc.gnu.org, amacleod@cygnus.com (Andrew Macleod)
Subject: Re: libstdc++/2568
Date: Fri, 25 May 2001 23:43:02 +0200 (MEST)

 Brad Lucier writes:
 
 > When configured with
 > 
 > env CC='cc -xarch=v9 -xildoff' ../configure --prefix=/pkgs/gcc-2.96 sparcv9-sun-solaris2.8 
 > 
 > the bootstrap for today's 3.0 sources fails after a good compare with
 
 I don't see any of this.
 
 > while there is a c++configs.h in
 > 
 > sparcv9-sun-solaris2.8/sparcv7/libstdc++-v3/include/bits/c++config.h
 > 
 > relative to my build directory.
 
 In a vanilla sparc-sun-solaris2.8 bootstrap for 3.0 (only
 
 	configure --prefix=/vol/gcc --local-prefix=/vol/gcc
 
 with gcc 2.95 as bootstrap compiler), I have
 
 	sparc-sun-solaris2.8/libstdc++-v3/include/bits/c++config.h
 
 Similarly for sparcv9-sun-solaris2.8, bootstrapped with
 
 	CC='cc -xarch=v9' \
 	configure --prefix=/vol/gcc --with-local-prefix=/vol/gcc \
 		sparcv9-sun-solaris2.8
 
 	sparcv9-sun-solaris2.8/libstdc++-v3/include/bits/c++config.h
 	sparcv9-sun-solaris2.8/sparcv7/libstdc++-v3/include/bits/c++config.h
 
 > This appears to be a configure problem; I already reported this problem in
 > 
 > http://gcc.gnu.org/ml/gcc-patches/2001-05/msg01430.html
 
 This report seems to indicate that you have Alexandre's patch to re-enable
 32x64 bi-arch support in your tree?  I'm just now running the bootstrap
 with this, but it's not there yet.
 
 	Rainer
 
 -----------------------------------------------------------------------------
 Rainer Orth, Faculty of Technology, Bielefeld University
 
 Email: ro@TechFak.Uni-Bielefeld.DE

Re: libstdc++/2568

[Brad Lucier]

The following reply was made to PR libstdc++/2568; it has been noted by GNATS.

From: Brad Lucier <lucier@math.purdue.edu>
To: ro@TechFak.Uni-Bielefeld.DE (Rainer Orth)
Cc: lucier@math.purdue.edu (Brad Lucier), bkoz@gcc.gnu.org,
   gcc-gnats@gcc.gnu.org, David.Billinghurst@riotinto.com, neil@gcc.gnu.org,
   amacleod@cygnus.com (Andrew Macleod)
Subject: Re: libstdc++/2568
Date: Fri, 25 May 2001 16:19:53 -0500 (EST)

 When configured with
 
 env CC='cc -xarch=v9 -xildoff' ../configure --prefix=/pkgs/gcc-2.96 sparcv9-sun-solaris2.8 
 
 the bootstrap for today's 3.0 sources fails after a good compare with
 
 /bin/sh ../libtool --mode=compile /export/home/lucier/gcc/gcc-3.0/objdir/gcc/xgcc -B/export/home/lucier/gcc/gcc-3.0/objdir/gcc/ -B/pkgs/gcc-2.96/sparcv9-sun-solaris2.8/bin/ -B/pkgs/gcc-2.96/sparcv9-sun-solaris2.8/lib/ -isystem /pkgs/gcc-2.96/sparcv9-sun-solaris2.8/include -DHAVE_CONFIG_H -I. -I../../../../libstdc++-v3/libmath -I.. -I../../../../libstdc++-v3/include -I../include  	     -g -c ../../../../libstdc++-v3/libmath/signbit.c
 mkdir .libs
 /export/home/lucier/gcc/gcc-3.0/objdir/gcc/xgcc -B/export/home/lucier/gcc/gcc-3.0/objdir/gcc/ -B/pkgs/gcc-2.96/sparcv9-sun-solaris2.8/bin/ -B/pkgs/gcc-2.96/sparcv9-sun-solaris2.8/lib/ -isystem /pkgs/gcc-2.96/sparcv9-sun-solaris2.8/include -DHAVE_CONFIG_H -I. -I../../../../libstdc++-v3/libmath -I.. -I../../../../libstdc++-v3/include -I../include -g -c ../../../../libstdc++-v3/libmath/signbit.c  -fPIC -DPIC -o .libs/signbit.o
 In file included from ../../../../libstdc++-v3/libmath/signbit.c:32:
 ../../../../libstdc++-v3/libmath/mathconf.h:31:28: bits/c++config.h: No such file or directory
 
 while there is a c++configs.h in
 
 sparcv9-sun-solaris2.8/sparcv7/libstdc++-v3/include/bits/c++config.h
 
 relative to my build directory.
 
 This appears to be a configure problem; I already reported this problem in
 
 http://gcc.gnu.org/ml/gcc-patches/2001-05/msg01430.html
 
 Brad

Re: libstdc++/2568

[Brad Lucier]

The following reply was made to PR libstdc++/2568; it has been noted by GNATS.

From: Brad Lucier <lucier@math.purdue.edu>
To: ro@TechFak.Uni-Bielefeld.DE (Rainer Orth)
Cc: lucier@math.purdue.edu (Brad Lucier), bkoz@gcc.gnu.org,
   gcc-gnats@gcc.gnu.org, David.Billinghurst@riotinto.com, neil@gcc.gnu.org,
   amacleod@cygnus.com (Andrew Macleod)
Subject: Re: libstdc++/2568
Date: Fri, 25 May 2001 14:32:40 -0500 (EST)

 Rainer Orth wrote:
 
 > I'm not sure what I'll find, but I will compare make check results of
 > 
 > * a sparc-sun-solaris2.8 configuration (sparcv7),
 
 I just did one of these today, and I started a sparcv9-sun-solaris2.8
 with (I believe) multilib support that should finish in a few
 hours.  (The bootstrap is fairly fast, it's the make check
 that kills me.)
 
 Brad

Re: libstdc++/2568

[Rainer Orth]

The following reply was made to PR libstdc++/2568; it has been noted by GNATS.

From: Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
To: Brad Lucier <lucier@math.purdue.edu>
Cc: bkoz@gcc.gnu.org, gcc-gnats@gcc.gnu.org, David.Billinghurst@riotinto.com,
        neil@gcc.gnu.org, amacleod@cygnus.com (Andrew Macleod)
Subject: Re: libstdc++/2568
Date: Fri, 25 May 2001 21:28:14 +0200 (MEST)

 Brad Lucier writes:
 
 > Is this on the 3.0 branch or on the trunk (3.1)?  I don't care if 3.0
 
 The 3.0 branch.  I don't start working on the trunk until the 3.0 release
 is out of the door; I'd like to make sure as many problems as possible are
 fixed once GCC 3.0 goes out.
 
 > bootstraps on sparcv9 because I don't believe that 64-bit support
 > works on the branch because Jakub's subreg-byte patches have not been applied
 > there, and I can't find any code that looks like this in c-lex.c on the
 > trunk at all, i.e., I couldn't even apply this patch manually to the trunk.
 
 I'm not sure what I'll find, but I will compare make check results of
 
 * a sparc-sun-solaris2.8 configuration (sparcv7),
 
 * sparcv9-sun-solaris2.8 without multilib options (i.e. -m64) (sparcv9)
 
 * sparcv9-sun-solaris2.8 with -m32 (sparcv7)
 
 and maybe, time permitting, though this is by far the most interesting to
 me
 
 * bi-arch sparc-sun-solaris2.8 (default -m32) (sparcv7)
 
 * bi-arch sparc-sun-solaris2.8 -m64 (sparcv9)
 
 once I've re-enabled bi-arch/multilibbing for this configuration).
 
 	Rainer
 
 -----------------------------------------------------------------------------
 Rainer Orth, Faculty of Technology, Bielefeld University
 
 Email: ro@TechFak.Uni-Bielefeld.DE

Re: libstdc++/2568

[Brad Lucier]

The following reply was made to PR libstdc++/2568; it has been noted by GNATS.

From: Brad Lucier <lucier@math.purdue.edu>
To: ro@TechFak.Uni-Bielefeld.DE (Rainer Orth)
Cc: bkoz@gcc.gnu.org, gcc-gnats@gcc.gnu.org, lucier@math.purdue.edu,
   David.Billinghurst@riotinto.com, neil@gcc.gnu.org,
   amacleod@cygnus.com (Andrew Macleod)
Subject: Re: libstdc++/2568
Date: Fri, 25 May 2001 14:16:25 -0500 (EST)

 Rainer Orth wrote:
 > Applying Andrew's patch
 > 
 > 	http://gcc.gnu.org/ml/gcc-patches/2001-04/msg01155.html
 > 
 > allowed the bootstrap to complete, and the make check is now on it's way.
 
 Is this on the 3.0 branch or on the trunk (3.1)?  I don't care if 3.0
 bootstraps on sparcv9 because I don't believe that 64-bit support
 works on the branch because Jakub's subreg-byte patches have not been applied
 there, and I can't find any code that looks like this in c-lex.c on the
 trunk at all, i.e., I couldn't even apply this patch manually to the trunk.
 
 Brad

Re: libstdc++/2568

[Rainer Orth]

The following reply was made to PR libstdc++/2568; it has been noted by GNATS.

From: Rainer Orth <ro@TechFak.Uni-Bielefeld.DE>
To: bkoz@gcc.gnu.org, gcc-gnats@gcc.gnu.org, lucier@math.purdue.edu,
        David.Billinghurst@riotinto.com, neil@gcc.gnu.org,
        Andrew Macleod <amacleod@cygnus.com>
Cc:  
Subject: Re: libstdc++/2568
Date: Fri, 25 May 2001 21:04:25 +0200 (MEST)

 I've just run into the same problem bootstrapping for
 sparcv9-sun-solaris2.8:
 
 /vol/gcc/obj/gcc-3.0-20010523/8-sparcv9-cc.new/gcc/xgcc -B/vol/gcc/obj/gcc-3.0-20010523/8-sparcv9-cc.new/gcc/ -nostdinc++ -L/vol/gcc/obj/gcc-3.0-20010523/8-sparcv9-cc.new/sparcv9-sun-solaris2.8/libstdc++-v3/src -L/vol/gcc/obj/gcc-3.0-20010523/8-sparcv9-cc.new/sparcv9-sun-solaris2.8/libstdc++-v3/src/.libs -B/vol/gcc/sparcv9-sun-solaris2.8/bin/ -B/vol/gcc/sparcv9-sun-solaris2.8/lib/ -isystem /vol/gcc/sparcv9-sun-solaris2.8/include -nostdinc++ -I/vol/gnu/src/gcc/gcc-3.0-branch-dist/libstdc++-v3/include -I/vol /gnu/src/gcc/gcc-3.0-branch-dist/libstdc++-v3/include/std -I/vol/gnu/src/gcc/gcc-3.0-branch-dist/libstdc++-v3/include/c_std -I../include -I/vol/gnu/src/gcc/gcc-3.0-branch-dist/libstdc++-v3/libsupc++ -I../libio -I/vol/gnu/src/gcc/gcc-3.0-branch-dist/libstdc++-v3/libio -I/vol/gnu/src/gcc/gcc-3.0-branch-dist/libstdc++-v3/libmath -g -O2 -fno-implicit-templates -Wall -Wno-format -W -Wwrite-strings -Winline -fdiagnostics-show-location=once -ffunction-sections -fdata-sections -g -!
 c c++locale.cc  -fPIC -DPIC -o .libs/c++locale.o
 c++locale.cc: In member function `void 
    std::numpunct<_CharT>::_M_initialize_numpunct(int*) [with _CharT = 
    wchar_t]':
 c++locale.cc:66: character constant too long
 c++locale.cc:67: character constant too long
 c++locale.cc: In member function `void std::moneypunct<_CharT, 
    _Intl>::_M_initialize_moneypunct(int*) [with _CharT = wchar_t, bool _Intl = 
    false]':
 c++locale.cc:96: character constant too long
 c++locale.cc:97: character constant too long
 make[4]: *** [c++locale.lo] Error 1
 make[4]: Leaving directory `/vol/gcc/obj/gcc-3.0-20010523/8-sparcv9-cc.new/sparcv9-sun-solaris2.8/libstdc++-v3/src'
 
 Applying Andrew's patch
 
 	http://gcc.gnu.org/ml/gcc-patches/2001-04/msg01155.html
 
 allowed the bootstrap to complete, and the make check is now on it's way.
 
 This is most likely the same problem as c/1739, so I'm Cc'ing David and
 Neil on this as well.  A make check on IRIX 6.2 with -mabi=64 is just on
 it's way (without the c-lex.c fix above applied), and I'll probably
 re-bootstrap when I'm done with that one.  Unfortunately, the machine (an 2
 CPU R8000 Power Challenge/L) is terribly slow, so this will most likely
 take a considerable amount of time.
 
 Regards.
 
 	Rainer
 
 -----------------------------------------------------------------------------
 Rainer Orth, Faculty of Technology, Bielefeld University
 
 Email: ro@TechFak.Uni-Bielefeld.DE

Re: libstdc++/2568

[Brad Lucier]

The following reply was made to PR libstdc++/2568; it has been noted by GNATS.

From: Brad Lucier <lucier@math.purdue.edu>
To: bkoz@gcc.gnu.org
Cc: gcc-gnats@gcc.gnu.org, lucier@math.purdue.edu, nobody@gcc.gnu.org
Subject: Re: libstdc++/2568
Date: Wed, 9 May 2001 02:39:58 -0500 (EST)

 The problem is still there as of today's CVS sources.
 
 Brad

Re: libstdc++/2568

[bkoz]

The following reply was made to PR libstdc++/2568; it has been noted by GNATS.

From: bkoz@gcc.gnu.org
To: bkoz@gcc.gnu.org, gcc-gnats@gcc.gnu.org, lucier@math.purdue.edu,
  nobody@gcc.gnu.org
Cc:  
Subject: Re: libstdc++/2568
Date: 9 May 2001 06:52:11 -0000

 Synopsis: libstdc++ build failure on sparcv9: "character constant too long"
 
 Responsible-Changed-From-To: unassigned->bkoz
 Responsible-Changed-By: bkoz
 Responsible-Changed-When: Tue May  8 23:52:11 2001
 Responsible-Changed-Why:
     Responsible.
 State-Changed-From-To: open->analyzed
 State-Changed-By: bkoz
 State-Changed-When: Tue May  8 23:52:11 2001
 State-Changed-Why:
     Brad what is current status on this?
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=2568&database=gcc